**3.3. Biomagnification factors**

The interpretation of the data resulting from the use of biomagnification factors are focused on BMF**TL** as the BMF and BMF**TL\*** was used in this study as an optional approach for evaluation of BMF methods. When the BMF is calculated for the Galapagos sea lion/thread herring case, the BMF values were consistent among the methodologies used (Table 2). In contrast, the three methods differed markedly from 9 to 9.5x1018 orders of magnitude higher for OC pesticides and from 4.8 to 1.9 x107 orders of magnitude higher for PCBs when the predator-prey BMFTL approaches versus the conventional C*PREDATOR*/C*PREY* ratio in the Galapagos sea lion/mullet relationship are compared. These fluctuations appear to be driven by the effect of the magnitude resulting from the differences in trophic levels. While the trophic level difference (TL predator − TLprey = 1.1) between the Galapagos sea lion and the thread herring is large, the trophic level difference (TL predator − TLprey = 0.11) between the Galapagos sea lion and the mullet is statistically insignificant (*p* >0.05) and cannot be used in the calculation of the predator-prey BMFTL .Thus, the predator-prey biomagnification factor methodologies (BMFTL) are sensitive to small differences in trophic levels (i.e., Galapagos sea lion-mullet). Based on this observation, the best way of expressing the BMF is the calculation of the BMF calculated as the C*PREDATOR*/C*PREY* ratio**,** which was similar between the Galapagos sea lion/herring and Galapagos sea lion/mullet cases.The use of different biomagnification factor measures showed that BMFTL and BMFTL\* are more appropriate to assess biomagnification if differences in trophic levels of predator/prey relationships are large (i.e. >1), as depicted in Table 2.

Calculated biomagnification factors of OC pesticides and PCB congeners, including octanolwater (KOW) and octanol-air partition coefficients (KOA), are shown in Table 2. The BMFTL of OC pesticides ranged from 7.3 (*trans*-chlordane) to 140 (*p*,*p*'-DDT) kg/kg lipid in Galapagos sea lion/thread herring and from 130 (*trans*-chlordane) to as high as 2000 (*p*,*p*'-DDE) kg/kg lipid in Galapagos sea lion/mullet, while BMFTL for PCB congeners ranged from 2.7 (PCB 156) to 30 (PCB 74) kg/kg lipid in Galapagos sea lion/thread herring, and from 11 (PCB 52) to 72 (PCB 153) kg/kg lipid in Galapagos sea lion/mullet (Table 3). No significant correlations were found between the BMFTL of OC pesticides and KOW (Figure 4b,d). Yet, BMFTL values decrease for some pesticides (e. g., mirex; trans-chlordane) when a KOW of 105.5 or 106.0 is exceeded. As a function of the octanol-air partition coefficient (KOA), the BMFTL for OC pesticides increased markedly as the KOA increased from 107.5 to 109, and then dropped for the rest of pesticides as KOA exceeds 109.5 (Figure 4a,c).

Assessing Biomagnification and Trophic Transport of Persistent Organic Pollutants in the Food Chain of the Galapagos Sea Lion (*Zalophus wollebaeki*): Conservation and Management Implications 93


NR= non reported;

92 New Approaches to the Study of Marine Mammals

The relative concentrations of contaminants observed in all sites exhibited a general common pattern, ∑DDT > ∑Chlordane > ∑PCBs >*β*-HCH> dieldrin > mirex, which was dominated by ∑DDTs, followed by chlordanes and PCBs, and secondly by *β*-HCH, dieldrin and mirex. Concentrations of ∑PCBs and OC pesticides detected in Galapagos sea lion pups showed no significant differences among rookeries (ANOVA for all comparisons, *p*> 0.05). This might suggest a common, global source of contamination delivering POPs to the

The interpretation of the data resulting from the use of biomagnification factors are focused on BMF**TL** as the BMF and BMF**TL\*** was used in this study as an optional approach for evaluation of BMF methods. When the BMF is calculated for the Galapagos sea lion/thread herring case, the BMF values were consistent among the methodologies used (Table 2). In contrast, the three methods differed markedly from 9 to 9.5x1018 orders of magnitude higher for OC pesticides and from 4.8 to 1.9 x107 orders of magnitude higher for PCBs when the predator-prey BMFTL approaches versus the conventional C*PREDATOR*/C*PREY* ratio in the Galapagos sea lion/mullet relationship are compared. These fluctuations appear to be driven by the effect of the magnitude resulting from the differences in trophic levels. While the trophic level difference (TL predator − TLprey = 1.1) between the Galapagos sea lion and the thread herring is large, the trophic level difference (TL predator − TLprey = 0.11) between the Galapagos sea lion and the mullet is statistically insignificant (*p* >0.05) and cannot be used in the calculation of the predator-prey BMFTL .Thus, the predator-prey biomagnification factor methodologies (BMFTL) are sensitive to small differences in trophic levels (i.e., Galapagos sea lion-mullet). Based on this observation, the best way of expressing the BMF is the calculation of the BMF calculated as the C*PREDATOR*/C*PREY* ratio**,** which was similar between the Galapagos sea lion/herring and Galapagos sea lion/mullet cases.The use of different biomagnification factor measures showed that BMFTL and BMFTL\* are more appropriate to assess biomagnification if differences in

animals, and that localized sources play a little role in contributions of POPs.

trophic levels of predator/prey relationships are large (i.e. >1), as depicted in Table 2.

the rest of pesticides as KOA exceeds 109.5 (Figure 4a,c).

Calculated biomagnification factors of OC pesticides and PCB congeners, including octanolwater (KOW) and octanol-air partition coefficients (KOA), are shown in Table 2. The BMFTL of OC pesticides ranged from 7.3 (*trans*-chlordane) to 140 (*p*,*p*'-DDT) kg/kg lipid in Galapagos sea lion/thread herring and from 130 (*trans*-chlordane) to as high as 2000 (*p*,*p*'-DDE) kg/kg lipid in Galapagos sea lion/mullet, while BMFTL for PCB congeners ranged from 2.7 (PCB 156) to 30 (PCB 74) kg/kg lipid in Galapagos sea lion/thread herring, and from 11 (PCB 52) to 72 (PCB 153) kg/kg lipid in Galapagos sea lion/mullet (Table 3). No significant correlations were found between the BMFTL of OC pesticides and KOW (Figure 4b,d). Yet, BMFTL values decrease for some pesticides (e. g., mirex; trans-chlordane) when a KOW of 105.5 or 106.0 is exceeded. As a function of the octanol-air partition coefficient (KOA), the BMFTL for OC pesticides increased markedly as the KOA increased from 107.5 to 109, and then dropped for

*3.2.3. Intersite comparisons* 

**3.3. Biomagnification factors** 

Values for log KOW and log KOA were obtained from Kelly *et al.* [2] and Mackay *et al*. [56].

**Table 2.** Biomagnification factors (BMF), Predator-prey Biomagnification factors (BMF*TL*) and Log Predator-prey Biomagnification factors (BMF*TL*\*) in units of kg/kg lipid for organochlorine pesticides (OCP) and PCB congeners in the Galapagos sea lion. The logarithmic values of the octanol-water (KOW) and octanol-air (KOA) partition coefficients for each contaminant are also reported as supporting indicators of bioaccumulation.

Assessing Biomagnification and Trophic Transport of Persistent Organic Pollutants in the Food Chain of the Galapagos Sea Lion (*Zalophus wollebaeki*): Conservation and Management Implications 95

**1**

**100**

**1**

**10**

**BMFTL (PCBs)**

**Figure 5.** Predator-prey biomagnification factors (BMFTL) in the Galapagos sea lion as expressed by the PCB congeners' concentration ratios sea lion/mullet (a, b) and sea lion/thread herring (c, d) as a function of log KOA (a, c) and log KOW (b, d). For PCBs, log KOW appears to be an adequate predictor of the bioaccumulative potential of PCBs in marine mammals because all PCBs tested have a high log KOA> 6.

The BMFTL for organochlorine pesticides expressed by the concentration ratios sea lion/thread herring and sea lion/mullet of the Galapagos sea lion are higher than those reported for harp seals (*Pagophilus groenlandicus*) from the contaminated Barents Sea [15], (Table 3). However, the BMFTL for PCBs of the Galapagos sea lion are lower than those reported for harp seals. This indicates the biomagnification predominance of organochlorine pesticides in tropical-equatorial regions versus the predominant biomagnification of PCBs in Arctic regions. To further explore these comparisons, the ratio of the BMFTL for *p*,*p*'-DDE (the DDT dominant metabolite) to the BMFTL for PCB 153 (used here as the most recalcitrant PCB congener) was calculated for both species of pinnipeds and then compared. As shown in Table 3, the ratio *p*,*p*'-DDE BMFTL/PCB 153 BMFTL was much higher in the Galapagos compared to that of the Barents Sea, which is driven by the predominance of *p*,*p*'-DDE biomagnification in the Galapagos. Vapor pressures of organic contaminants are expected to be higher in tropical systems due to warmer/higher temperature in comparisons to cold/lower temperature in the Arctic; and, therefore, higher thermodynamic gradients and increase in concentrations are likely to occur during the trophic transfer of contaminant

**10**

**BMFTL (PCBs)**

**Galapagos sea lion/mullet**

**74**

**a b**

**118**

**95**

**7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 Log KOA**

**c d**

**95**

**7.0 8.0 9.0 10.0 11.0 Log KOA**

**118 128**

**99 <sup>101</sup> <sup>105</sup>**

**<sup>138</sup> <sup>146</sup> 153**

**201**

Values for log KOW and log KOA were obtained from Kelly *et al.* [2] and Mackay *et al*. [56].

mass from prey to predator, resulting in a high biomagnification factor.

**74**

**128**

**99 <sup>101</sup> <sup>105</sup>**

**146**

**138**

**153**

**180 183 <sup>187</sup> <sup>201</sup>**

**156 174**

**52**

**Galapagos sea lion/thread herring**

**52**

**1**

**1**

**10**

**BMFTL (PCBs)**

**100**

**10**

**BMFTL (PCBs)**

**100**

**100**

**Galapagos sea lion/mullet**

**99**

**105 <sup>118</sup> <sup>128</sup>**

**5.5 6.0 6.5 7.0 7.5 8.0 Log KOW**

**99**

**105 118 128 138**

**5.5 6.0 6.5 7.0 7.5 8.0 Log KOW**

**153**

**174 183 180 187 201 202**

**101**

**153**

**101**

**Galapagos sea lion/thread herring**

**52**

**74**

**74**

**95**

**156**

**146**

**95**

**<sup>156</sup> <sup>174</sup>**

**138 146**

**52**

**Figure 4.** Predator-prey biomagnification factors (BMFTL) in the Galapagos sea lion as expressed by the OC pesticide concentration ratios sea lion/ mullet (a, b) and sea lion/ thread herring (c, d) as a function of log KOA (a, c) and log KOW (b, d).The figure illustrates that while the Stockholm Convention for POPs uses a log KOW> 5 as a criterion to identify bioaccumulative substances, substances including *β*-HCH with a log KOW< 5 can biomagnify in marine mammals. Log KOA appears to be a better predictor of substances that have the potential to biomagnify in marine mammals. Values for log KOW and log KOA were obtained from Kelly *et al.* [2] and Mackay *et al*. [56].

The BMFTL of PCBs showed different trends when looking a different prey items in terms of KOA. While no correlation was found between the BMFTL of PCBs and log KOA in the Galapagos sea lion/ mullet relationship (Figure 5a), BMFTL for PCBs increased as the KOA increased from 107.6 to 108.4 and then appeared to decrease gradually with increasing log KOA in the Galapagos sea lion/thread herring relationship (Figures 5c). No correlation was found between the BMFTL of PCBs and log KOW for the Galapagos sea lion/thread herring or Galapagos sea lion/mullet feeding relationship (Figure 5b, d).

These observations demonstrate that these halogenated substances biomagnify and achieve concentrations in Galapagos sea lions that exceed those in their prey, although physiological processes and biotransformation may limit the biomagnification of some contaminants. When comparing the plots of BMFTL of PCBs versus log KOW or versus log KOA similar patterns were observed for both Galapagos sea lion/thread herring and Galapagos sea lion/mullet feeding relationships (Figure 5a,d and Figure 5b,d, respectively). This is explained by the strong correlation usually observed between log KOA and log KOW of PCBs [53].

*p***,***p***'-DDE**

**dieldrin**

*cis* **-nonachlor mirex**

**a b**

**5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 Log KOA**

**c d**

**mirex** *cis* **-nonachlor**

**5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 Log KOA**

were obtained from Kelly *et al.* [2] and Mackay *et al*. [56].

Galapagos sea lion/mullet feeding relationship (Figure 5b, d).

correlation usually observed between log KOA and log KOW of PCBs [53].

**dieldrin**

*p***,***p***'-DDD**

*cis* **-chlordane**

*p***,***p***'-DDT**

*β* **-HCH** *trans* **-chlordane**

*trans* **-nonachlor**

*trans* **-nonachlor**

*p***,***p***'-DDE** *p***,***p***'-DDT**

*cis* **-chlordane**

*trans* **-chlordane**

*β* **-HCH** *p***,***p***'-DDD**

**Galapagos sea lion/thread herring**

**Galapagos sea lion/mullet**

*β-* **HCH**

*β-* **HCH**

**1**

**10**

**100**

**BMFTL (OCP)**

**Figure 4.** Predator-prey biomagnification factors (BMFTL) in the Galapagos sea lion as expressed by the OC pesticide concentration ratios sea lion/ mullet (a, b) and sea lion/ thread herring (c, d) as a function of log KOA (a, c) and log KOW (b, d).The figure illustrates that while the Stockholm Convention for POPs uses a log KOW> 5 as a criterion to identify bioaccumulative substances, substances including *β*-HCH with a log KOW< 5 can biomagnify in marine mammals. Log KOA appears to be a better predictor of substances that have the potential to biomagnify in marine mammals. Values for log KOW and log KOA

The BMFTL of PCBs showed different trends when looking a different prey items in terms of KOA. While no correlation was found between the BMFTL of PCBs and log KOA in the Galapagos sea lion/ mullet relationship (Figure 5a), BMFTL for PCBs increased as the KOA increased from 107.6 to 108.4 and then appeared to decrease gradually with increasing log KOA in the Galapagos sea lion/thread herring relationship (Figures 5c). No correlation was found between the BMFTL of PCBs and log KOW for the Galapagos sea lion/thread herring or

These observations demonstrate that these halogenated substances biomagnify and achieve concentrations in Galapagos sea lions that exceed those in their prey, although physiological processes and biotransformation may limit the biomagnification of some contaminants. When comparing the plots of BMFTL of PCBs versus log KOW or versus log KOA similar patterns were observed for both Galapagos sea lion/thread herring and Galapagos sea lion/mullet feeding relationships (Figure 5a,d and Figure 5b,d, respectively). This is explained by the strong

**1000**

**10**

**100**

**1000**

**BMFTL (OCP)**

**10000**

**mirex**

*cis-* **nonachlor** *trans* -**nonachlor**

**dieldrin**

**dieldrin**

**3.5 4.5 5.5 6.5 7.5 8.5 Log KOW**

*trans* **-nonachlor** *cis-* **nonachlor**

*cis-* **chlordane**

**3.5 4.5 5.5 6.5 7.5 8.5 Log KOW**

*p*,*p*'-**DDD**

*p,p'-* **DDT** *p,p'-* **DDE**

*p,p'-* **DDE**

**mirex**

*p*,*p* '-**DDT** *p,p'-* **DDD**

*cis* **-chlordane**

*trans-c***hlordane**

*trans-* **chlordane**

**Galapagos sea lion/mullet**

**Galapagos sea lion/thread herring**

**1**

**1**

**10**

**100**

**BMFTL (OCP)**

**1000**

**10**

**100**

**BMFTL (OCP)**

**1000**

**10000**

**Figure 5.** Predator-prey biomagnification factors (BMFTL) in the Galapagos sea lion as expressed by the PCB congeners' concentration ratios sea lion/mullet (a, b) and sea lion/thread herring (c, d) as a function of log KOA (a, c) and log KOW (b, d). For PCBs, log KOW appears to be an adequate predictor of the bioaccumulative potential of PCBs in marine mammals because all PCBs tested have a high log KOA> 6. Values for log KOW and log KOA were obtained from Kelly *et al.* [2] and Mackay *et al*. [56].

The BMFTL for organochlorine pesticides expressed by the concentration ratios sea lion/thread herring and sea lion/mullet of the Galapagos sea lion are higher than those reported for harp seals (*Pagophilus groenlandicus*) from the contaminated Barents Sea [15], (Table 3). However, the BMFTL for PCBs of the Galapagos sea lion are lower than those reported for harp seals. This indicates the biomagnification predominance of organochlorine pesticides in tropical-equatorial regions versus the predominant biomagnification of PCBs in Arctic regions. To further explore these comparisons, the ratio of the BMFTL for *p*,*p*'-DDE (the DDT dominant metabolite) to the BMFTL for PCB 153 (used here as the most recalcitrant PCB congener) was calculated for both species of pinnipeds and then compared. As shown in Table 3, the ratio *p*,*p*'-DDE BMFTL/PCB 153 BMFTL was much higher in the Galapagos compared to that of the Barents Sea, which is driven by the predominance of *p*,*p*'-DDE biomagnification in the Galapagos. Vapor pressures of organic contaminants are expected to be higher in tropical systems due to warmer/higher temperature in comparisons to cold/lower temperature in the Arctic; and, therefore, higher thermodynamic gradients and increase in concentrations are likely to occur during the trophic transfer of contaminant mass from prey to predator, resulting in a high biomagnification factor.


Assessing Biomagnification and Trophic Transport of Persistent Organic Pollutants in the Food Chain of the Galapagos Sea Lion (*Zalophus wollebaeki*): Conservation and Management Implications 97

calculated BMFs may not always reflect actual biomagnification [54]. As shown in this study, predator-prey BMFs revealed the biomagnification capacity of POPs in the food chain of the Galapagos sea lions, which is an apex predator possessing flexible feeding preferences

Efficient uptake and dietary assimilation and slow depuration/excretion rates of these compounds (PCBs with KOW ranging 105−107, and OC pesticides KOW ranging 103.8−107.0) explain the high degree of biomagnification in the Galapagos marine food chain. Dietary absorption efficiencies of Penta and Hexachlorobiphenyls are typically between 50-80% in fish and 90-100% in mammals [55] and chemical half-lives (*t*1/2) for recalcitrant PCBs such as PCB 153 in organisms exceed 1000 days [56]. The analysis of BMFTL estimates of PCBs and OC pesticides (Figures 4-5) indicates that OC pesticides and PCBs are accumulated by fish and sea lions and also biomagnify in the food chain. Based on contaminants' predator-prey BMFs, the DDT metabolites, *p*,*p*'-DDT and *p*,*p*'-DDE, followed by *trans*-nonachlor (Figure 4), are the most bioaccumulative pesticides, while PCB 74 and 153 are the most bioaccumulative PCB congeners in the Galapagos sea lion (Figures 5). The less

Of particular importance is the biomagnification behaviour of *β*-HCH with a KOW< 104 (KOW = 103.8; Figure 4b,d), but with a KOA of 108.9−1010.5 (Figure 4a,c), contrasting with the regulatory criteria and current management policies (i.e. Stockholm Convention; CEPA) for POPs that consider only chemicals with KOW values >105 as bioaccumulative substances [7]. The predatorprey biomagnification factors (BMFTL = 63−552) of *β*-HCH in Galapagos sea lions exceed equivalent biomagnification factors of PCB 153 (BMFTL =18.0−72.2) and PCB 74 (BMFTL =30.0−72.0), as shown in Table 2. This portrays that *β*-HCH, a relatively hydrophilic and nonmetabolizable chemical, biomagnifies in the tropical marine mammalian food chain of an air breathing organism (the Galapagos sea lions), which is explained by the relatively high KOAof *β*-HCH (KOA> 107.0) and its negligible respiratory elimination. Biomagnification of *β*-HCH was evident in the lichen-caribou-wolf terrestrial food chain, in the maritime and interior grizzly bears' food chains, and in a marine mammalian food web (including water-respiring and airbreathing organisms) from temperate regions of Canada and the Canadian Arctic [2,14,19].

Lack of significant differences and consistent uniformity of PCBs and OC pesticides, particularly for PCBs, among sites might indicate common sources of contamination. Concentrations of PCBs were also similar among rookeries in an earlier baseline study [34], although DDT concentrations were found to be significantly different [35]. Furthermore, principal components analysis represented a more comprehensive approach for exploring spatial differences and behaviour of POPs. The two first principal components (i.e., PC 1 and PC2) accounted for 55.2% of the total variation in Galapagos sea lion pups. PCA score plot results for the 2008 data further revealed that contaminants follow a similar trend, aggregated near to the centre of the axes, among sites, showing lack of discrimination and differentiation in contaminant patterns (Figure 6a). The first principal component (i.e.,

bioaccumulative compounds are *trans*-chlordane and PCB 156.

**3.5. Environmental transport of contaminants** 

(dietary plasticity).

NR= non reported

a Borga *et al*. [15].

**Table 3.** Comparison of BMFTL for remote marine food chains between the Galapagos Islands and an Arctic reion for selected organochlorine pesticides and PCBs. The BMFTL for Galapagos sea lions are expressed as the range of concentration ratios of both sea lion/thread herring and sea lion/mullet feeding relationships.
